Supplement to the paper "Part-of-speech tagging accuracy and its reference standard in a corpus of Saudi legal English" (Alhatlani, Muhammad; Alghizzi, Talal). The corpus the paper measures is the Saudi Legal English Corpus (SaLEC) version 2.0, a separate record at DOI 10.5281/zenodo.21328459, which also holds the tagging pipeline, its rule component and lexicons. Licence: Creative Commons Attribution 4.0 International (CC BY 4.0). The study is a blind double annotation of 2,000 tokens drawn from the corpus's four acquisition pools, adjudicated to a locked reference standard and scored under a design-weighted estimand with a document-clustered bootstrap; a validation of the pipeline's proper-noun correction rule on two judged rounds; and a query of the two largest Universal Dependencies English treebanks for the constructions where the pipeline and the reference diverge most. The evaluation measures part-of-speech assignment conditional on the pipeline's tokenization and sentence segmentation, since the sample was drawn from the pipeline's own token stream. This record holds the materials, the code, and the generated results behind every figure the paper reports from those three parts.